Data
Song mainly adopts largely-scale census, administrative, and representative household survey data. He has extensive experience working with cross-sectional and longitudinal datasets in the U.S. (e.g., ACS, Add Health, ATUS, BRFSS, NCHAT, NHIS, NSFG, NVSS) and China (e.g., CFPS, CGSS, WOMENSTATUS, 2016 Survey of the Fertility Decision-Making Processes in Chinese Families).
Song also has rich fieldwork experience, mostly in China. He co-designed and conducted original quantitative surveys in six Chinese provinces (2015-2017) and in-depth interviews in two Chinese cities (2018-2019). Currently, he is collecting original:
1) computational text data through social media (using Python)
2) experimental data through online platforms (using MTurk and Qualtrics)
3) qualitative data through fieldwork in China (using participant observation and in-depth interview)
3.1) one IRB approved PI project on divorced women
3.2) one IRB approved project on queer parents
Methodology
Song received interdisciplinary methods training in sociology, demography, economics, anthropology, spatial studies, and development studies. He also regularly attends cutting-edge short training programs. As a result, he is well-trained in:
- mainstream quantitative methods (e.g., OLS, LPM, logistic, probit, poisson regressions, log-linear models) [study 1] [study 2] [study 3] [study 4] [study 5]
- demographic methods & event history analyses (e.g., life tables, cox, and discrete-time models) [study 1]
- causal inference techniques (e.g., multilevel models, propensity score-based methods, heterogeneous effects, DID, RDD, IV) [study 2]
- qualitative methods (e.g., participant observation, in-depth interviews, Nvivo software)
- computational methods (e.g., machine learning techniques, topic modeling, dictionary-based sentiment analysis, LWIC)
- geospatial methods (e.g., GIS institute training certificate)
